Communication between a calling party (source) and a called party (destination) may be established over a communication network. Such a communication network may use routing protocols in order to establish connections over which such communication can occur. Communication networks that support source routing protocols typically comprise a number of individual communication switches through which calls are routed. A call set-up message is sent along a path through a number of intervening switches, or nodes, in order to establish the call.
In source routing protocols, each node in the network determines a path to the destination of a call based on current knowledge of that node of the network topology. A source node encodes the computed path in a message used to setup a given connection, so other nodes along the path can follow the computed path. As network topology changes (for example, nodes and links appear or disappear or bandwidth consumption changes) a path to a given destination may change. These changes are reflected in a path computed by the source node.
One type of topology change reported by nodes in the network is bandwidth available on each link. As the bandwidth on a given link changes, nodes at each end of the link report the new available bandwidth. Issuing an advertisement with each bandwidth change is usually not practical, as it would require significant resources to distribute the changes and to act upon them. These resources would often need to be taken from resources available for setup of calls, hence decreasing overall network efficiency. As a result, a concept of “significant change” is used: a node advertises a change in available bandwidth only when the bandwidth changes by a value deemed significant from the last advertised value. The significant-change-based advertisements have the following drawback: decreases in bandwidth that do not cross the significant value bound are deemed “insignificant”, and hence are not advertised to the network. The other nodes in the network, unaware of the decrease in bandwidth, keep using the last advertised bandwidth of that link in their path computations, even though the actual value is less than this last advertised value. Any of these nodes can build a path that includes the above link because its last advertised bandwidth satisfies the bandwidth requirements of a given call. However, if the bandwidth requirements of a call are greater than the bandwidth currently available on that link, the call blocks when a Call Admission Control rejects the request on a particular link.
Private Network-Network Interface (PNNI) protocol is an example of a source routing protocol that advertises bandwidth changes using a concept of the significant change. PNNI provides two control parameters that define what is deemed significant bandwidth change on a link: Available Cell Rate Proportional Multiplier (AvCR_PM) and Available Call Rate Minimum Threshold (AvCR_mT). AvCR_PM specifies the percentage that the bandwidth of the link must change from the last advertised value for a change to be deemed significant. AvCR_mT is a minimum threshold, expressed as a percentage of the maximum cell rate, ensuring a non-zero range of insignificance. As the bandwidth on the link is being consumed AvCR_PM is used by the node to determine significant changes until the value based on AvCR_mT (i.e. link bandwidth*AvCR_mT) becomes bigger than that based on AvCR_PM. When this takes place no further advertisement is issued until the link's bandwidth reaches zero or increases by the AvCR_mT-based value. When nodes in the network include the link in their paths as a result of its last outdated available bandwidth advertisement, and a call blocks, because it requires more bandwidth than the link's current available bandwidth, the call is released or cranked back. RELEASE message may indicate that the call blocked because the bandwidth on the link was not available, and may include the current available bandwidth (AvCR) on the link that blocked the call. An alternate routing may take place to attempt to avoid the blocked link. Such behaviour has the following drawbacks:    1. A node launching the call performs a re-route for that call around the blocked link but, since no update to the link's available bandwidth for all service categories occurs some outdated link information is still used in routing subsequent calls. This means that an unnecessary load is presented to the network for calls that try to use the link but fail because of an inadequate bandwidth. It will be appreciated that outdated link information is used only if AvCR information is included for the link;    2. The overall call's setup latency increases as calls block and need to be alternate routed;    3. In an extreme case, a call may fail to be setup if each routing attempt experiences the above-described problem; and    4. Only a source node or at most nodes along the path are informed about new available bandwidth when the call fails. All other nodes still have the outdated information.
There is a need for a routing system to address aspects of shortcomings of the prior art signaling systems.